SLmetrics (version 0.3-4)

specificity.factor: Specificity

Description

A generic S3 function to compute the specificity score for a classification model. This function dispatches to S3 methods in specificity() and performs no input validation. If you supply NA values or vectors of unequal length (e.g. length(x) != length(y)), the underlying C++ code may trigger undefined behavior and crash your R session.

Defensive measures

Because specificity() operates on raw pointers, pointer-level faults (e.g. from NA or mismatched length) occur before any R-level error handling. Wrapping calls in try() or tryCatch() will not prevent R-session crashes.

To guard against this, wrap specificity() in a "safe" validator that checks for NA values and matching length, for example:

safe_specificity <- function(x, y, ...) {
  stopifnot(
    !anyNA(x), !anyNA(y),
    length(x) == length(y)
  )
  specificity(x, y, ...)
}

Apply the same pattern to any custom metric functions to ensure input sanity before calling the underlying C++ code.

Efficient multi-metric evaluation

For multiple performance evaluations of a classification model, first compute the confusion matrix once via cmatrix(). All other performance metrics can then be derived from this one object via S3 dispatching:

## compute confusion matrix
confusion_matrix <- cmatrix(actual, predicted)
## evaluate specificity
## via S3 dispatching
specificity(confusion_matrix)
## additional performance metrics
## below

The specificity.factor() method calls cmatrix() internally, so explicitly invoking specificity.cmatrix() yourself avoids duplicate computation, yielding significant speed and memory effciency gains when you need multiple evaluation metrics.

Usage

# S3 method for factor
specificity(actual, predicted, estimator = 0L, na.rm = TRUE, ...)

Value

If estimator is given as

• 0 - a named <double> vector of length k

• 1 - a <double> value (Micro averaged metric)

• 2 - a <double> value (Macro averaged metric)

Arguments

actual, predicted

A pair of <integer> or <factor> vectors of length \(n\), and \(k\) levels.

estimator

An <integer>-value of length \(1\) (default: \(0\)).

• 0 - a named <double>-vector of length k (class-wise)

• 1 - a <double> value (Micro averaged metric)

• 2 - a <double> value (Macro averaged metric)

na.rm

A <logical> value of length \(1\) (default: TRUE). If TRUE, NA values are removed from the computation. This argument is only relevant when micro != NULL. When na.rm = TRUE, the computation corresponds to sum(c(1, 2, NA), na.rm = TRUE) / length(na.omit(c(1, 2, NA))). When na.rm = FALSE, the computation corresponds to sum(c(1, 2, NA), na.rm = TRUE) / length(c(1, 2, NA)).

...

Arguments passed into other methods.

Other names

The specificity has other names depending on research field:

• True Negative Rate, tnr()

• Selectivity, selectivity()

References

James, Gareth, et al. An introduction to statistical learning. Vol. 112. No. 1. New York: springer, 2013.

Hastie, Trevor. "The elements of statistical learning: data mining, inference, and prediction." (2009).

Pedregosa, Fabian, et al. "Scikit-learn: Machine learning in Python." the Journal of machine Learning research 12 (2011): 2825-2830.

See Also

Other Classification: accuracy(), auc.pr.curve(), auc.roc.curve(), baccuracy(), brier.score(), ckappa(), cmatrix(), cross.entropy(), dor(), fbeta(), fdr(), fer(), fmi(), fpr(), hammingloss(), jaccard(), logloss(), mcc(), nlr(), npv(), plr(), pr.curve(), precision(), recall(), relative.entropy(), roc.curve(), shannon.entropy(), zerooneloss()

Other Supervised Learning: accuracy(), auc.pr.curve(), auc.roc.curve(), baccuracy(), brier.score(), ccc(), ckappa(), cmatrix(), cross.entropy(), deviance.gamma(), deviance.poisson(), deviance.tweedie(), dor(), fbeta(), fdr(), fer(), fmi(), fpr(), gmse(), hammingloss(), huberloss(), jaccard(), logloss(), maape(), mae(), mape(), mcc(), mpe(), mse(), nlr(), npv(), pinball(), plr(), pr.curve(), precision(), rae(), recall(), relative.entropy(), rmse(), rmsle(), roc.curve(), rrmse(), rrse(), rsq(), shannon.entropy(), smape(), zerooneloss()

Examples

## Classes and
## seed
set.seed(1903)
classes <- c("Kebab", "Falafel")

## Generate actual
## and predicted classes
actual_classes <- factor(
    x = sample(x = classes, size = 1e3, replace = TRUE),
    levels = c("Kebab", "Falafel")
)

predicted_classes <- factor(
    x = sample(x = classes, size = 1e3, replace = TRUE),
    levels = c("Kebab", "Falafel")
)

## Evaluate performance
SLmetrics::specificity(
   actual    = actual_classes, 
   predicted = predicted_classes
)